470 JOURNAL OF CLIMATE VOLUME 19

NOTES AND CORRESPONDENCE

Impact of Great Anomalies on the Low-Frequency Variability of the North Atlantic Climate

RONG ZHANG AND GEOFFREY K. VALLIS Geophysical Fluid Dynamics Laboratory, and Atmospheric and Oceanic Sciences Program, Princeton University, Princeton, New Jersey

(Manuscript received 25 April 2005, in final form 4 August 2005)

ABSTRACT

In this paper, it is shown that coherent large-scale low-frequency variabilities in the North —that is, the variations of , deep western boundary current, northern recir- culation gyre, and Gulf Stream path—are associated with high-latitude oceanic Great Salinity Anomaly events. In particular, a dipolar sea surface temperature anomaly (warming off the U.S. east coast and cooling south of Greenland) can be triggered by the Great Salinity Anomaly events several years in advance, thus providing a degree of long-term predictability to the system. Diagnosed phase relationships among an observed proxy for Great Salinity Anomaly events, the sea surface temperature anomaly, and the North Atlantic Oscillation are also discussed.

1. Introduction ing the 1980s and 1990s (Belkin et al. 1998; Belkin 2004). There are also observational indications for the It has long been controversial as to whether the ob- occurrence of GSA events during the early twentieth served low-frequency variabilities in the North Atlantic century (Dickson et al. 1988; Walsh and Chapman 1990; Ocean, such as sea surface temperature (SST) anoma- Schmith and Hansen 2003). lies (Deser and Blackmon 1993; Kushnir 1994) and During GSA events, low-salinity water propagates Gulf Stream path shifts (Taylor and Stephens 1998; into Labrador Sea and reduces the deep water forma- Joyce et al. 2000) are directly forced by the atmospheric tion there (Lazier 1980). For example, the observed North Atlantic Oscillation (NAO; Hurrell 1995), as reductions of Labrador Seawater thickness during the suggested by Halliwell (1997) and Eden and Jung 1970s and 1980s (Curry et al. 1998) are related to GSA (2001), or whether the low-frequency NAO variability events. The reduction of Labrador Sea deep water for- is modulated by North Atlantic oceanic variability mation will lead to a weakening of the thermohaline (Rodwell et al. 1999; Mehta et al. 2000; Robertson et al. 2000; Latif 2001). A third option is that such variability circulation (THC) and deep western boundary current is caused by ocean–atmosphere coupled oscillations (DWBC; Fig. 1). Previous modeling studies showed (Delworth et al. 1993; Joyce et al. 2000). Here we show that the THC can be weakened by the negative polar that the North Atlantic low-frequency variations are surface salinity anomalies (Delworth et al. 1993; Häk- closely related to the Great Salinity Anomalies kinen 1993; Weaver et al. 1993). The modeling result of (GSAs). A strong GSA event occurred in the 1970s, Thompson and Schmitz (1989) showed that the Gulf when anomalous freshwater/sea ice from the Arctic Stream at the separation point is influenced by the ocean propagated into the North Atlantic subpolar gyre DWBC. However, this seems inconsistent with obser- (Dickson et al. 1988). The GSA events reoccurred dur- vations that there is little change in the separation point at Cape Hatteras on decadal time scales; rather, obser- vations show that major decadal shifts of the Gulf Corresponding author address: Rong Zhang, GFDL/AOS Pro- Stream path occurred in the open ocean downstream of gram, Princeton University, Princeton, NJ 08542. Cape Hatteras (Joyce et al. 2000). Zhang and Vallis E-mail: [email protected] (2005, manuscript submitted to J. Phys. Oceanogr.)

© 2006 American Meteorological Society Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

JCLI3623 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 471

(2000) did reproduce the results of Rodwell et al. (1999) with a different AGCM and showed that the correlation between simulated ensemble-mean NAO index (forced by observed SST anomaly) and observed NAO index is much higher for low-pass-filtered data than for unfiltered data. The ensemble averaging con- siderably reduces variability of the averaged sea level pressure (SLP) anomalies; hence, the simulated en- semble average NAO is considerably weaker than the observed NAO (Mehta et al. 2000). These AGCM simulations do suggest that the SST anomaly can modu- late the phase of NAO at low frequency, while the physical processes that control the amplitudes of NAO are not well simulated. Nevertheless, given the doubt as to the possible influence of the ocean on the atmo- sphere, we explore the influence of GSA events on North Atlantic climate variability with diagnoses of ob- FIG. 1. Schematic diagram of the mechanism: 1) propagation of served datasets of the twentieth century. GSA, 2) weakening of DWBC, and 3) weakening of NRG and a northward shift of mean Gulf Stream path. Here, “W” indicates warming and “C” indicates cooling. 2. Modeling the oceanic impacts of GSA events a. Model setup and experimental design showed that the bottom vortex stretching induced by the downslope DWBC over the steep continental slope We model impacts of GSAs on the low-frequency south of the Grand Banks influences the formation of variability in the North Atlantic with a 1° ϫ 1° global the observed cyclonic northern recirculation gyre ocean general circulation model (OGCM)—the Geo- (NRG) north of the Gulf Stream and keeps the Gulf physical Fluid Dynamics Laboratory (GFDL) Modular Stream path downstream of Cape Hatteras separated Ocean Model, version 4 (MOM4; Griffies et al. 2004) from the coast. Hence, as the weakened DWBC arrives coupled to an atmospheric energy balance model south of the Grand Banks a few years later, the bottom (EBM) with a hydrological cycle and a slab sea ice vortex stretching effect is weakened, resulting in the model (Winton 2000). This class of models provides a weakening of the NRG, a northward shift of Gulf useful and appropriate way to study the ocean’s low- Stream path, and warming off the U.S. east coast. frequency variability, free from the complications of Meanwhile the propagation of the low-salinity anomaly using a full atmospheric GCM, yet allowing large-scale in the subpolar gyre induces widespread cooling south atmospheric feedbacks to THC variations that are not of Greenland. Thus a dipolar SST anomaly appears in available with ocean-only models (e.g., Weaver et al. the North Atlantic. In this study, we simulate the above 2001). The coupled model is spun up for 51 yr with proposed processes with a numerical model. Comprehensive Ocean–Atmosphere Data Set Such phenomena may have important climate rami- (COADS; da Silva et al. 1994) climatological monthly fications. For example, Rodwell et al. (1999) showed mean surface wind stress and wind speed. The THC that positive phases of the NAO at decadal or longer (maximum meridional overturning streamfunction in time scales can be forced by a dipolar SST anomaly the North Atlantic) is stabilized after the spinup, al- (warming off the U.S. east coast and cooling south of though the spinup time is not long enough for the entire Greenland). We show that just such a dipolar SST deep ocean to reach equilibrium. This time scale is long anomaly could be triggered by GSA events that occur enough for the adjustment of DWBC in the western some years in advance and therefore that GSA events North Atlantic and upper-ocean properties in the could lead to changes in the phase of the NAO at very North Atlantic, which we focus on here. Gerdes and long time scales. However, there is still much debate as Köberle (1995) showed that the western boundary re- to whether the oceanic forcing of NAO found by Rod- gion in the North Atlantic deep ocean reaches near well et al. (1999) is a real effect—many modeling steady state within an adjustment period of 10 yr in studies with atmosphere general circulation models response to the surface buoyancy forcing in deep-water (AGCMs) show a weak response of NAO to extratrop- formation regions. This adjustment time scale of the ical SST forcing (Kushnir et al. 2002). Mehta et al. western boundary region in the North Atlantic deep

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 1 live 4/C 472 JOURNAL OF CLIMATE VOLUME 19

FIG. 2. Modeling results with all anomalous forcings (GW ϩ WIND ϩ GSA). (a) GSA index (color shade) and low-pass-filtered observed Iceland sea ice extent anomaly (green dashed line), normalized by their std devs. (b) Modeled (inverted) central Labrador Sea SST anomaly (K; red means cooling) and observed (inverted) anomaly (K; green dashed line; HadISST). (c) Modeled (inverted) DWBC anomaly at 44°N (Sv; red means less southward transport). (d) Modeled NRG anomaly [Sv; inverted, i.e., the anomaly of the barotropic streamfunction at the center of the NRG (41°N, 57°W); red means weaken- ing, less cyclonic]. (e) Modeled shifts of mean Gulf Stream path in degrees and that diagnosed from WOD98 (green dash line), i.e., the averaged location of the annual mean 15°C isotherm at 200 m between 75° and 55°W. ocean depends on the advection time scale of DWBC (Zhang and Vallis 2005, manuscript submitted to J. from deep-water formation source regions, not the Phys. Oceanogr.). background diffusive time scale. The coupled model is We simulate GSA events by adding a time series of integrated continuously for another 45 yr as a control anomalous freshwater flux over the past several de- experiment. The model produces a cyclonic NRG and a cades (Fig. 2a; “GSA index”) near the Greenland coast separated Gulf Stream path downstream of Cape Hat- in the model, assuming that it is proportional to ob- teras from the coast in the mean-state ocean circulation served SLP difference (COADS; da Silva et al. 1994)

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 2 live 4/C 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 473 between the Labrador Sea off the Greenland west coast TABLE 1. List of perturbed experiments. (63°N, 56°W) and the Arctic north of the Kara Sea GW ϩ (78°N, 75°E), smoothed with a 5-yr running mean. The GW ϩ WIND ϩ two locations were chosen because the observed SLP Anomalous external forcing GW WIND GSA GSA anomaly there has the maximum positive and negative Anomalous radiative forcing X X X covariance, respectively, with the Iceland sea ice extent Anomalous wind forcing X X anomaly (Fig. 2a) over the past several decades. The Anomalous freshwater forcing X X Iceland sea ice extent anomaly is derived from one of the long-term Icelandic sea ice records, the number of areas off the coasts of Iceland on which sea ice is ob- atmospheric greenhouse gases and COADS anomalous ϩ served in a sea ice year as defined and recorded in wind forcing (“GW WIND”), as well as the above- Wallevik and Sigurjónsson (1998). It is very similar to mentioned anomalous freshwater flux forcing due to another long-term Icelandic sea ice record, the well- GSA events (“GSA”), for a 45-yr period from January known Koch index (Kelly et al. 1987), which is defined 1946 to December 1990. The temporal variation of the as the number of weeks any sea ice is observed near the difference between each perturbed experiment and the coasts of Iceland in a sea ice year. The sea ice year is control experiment for the 45-yr period is calculated to defined in terms of the normal cycle of ice advance and represent the response to various anomalous external retreat, October–September, and is dated by the year of forcing, similar to the method used in Manabe et al. the January. These long-term Icelandic sea ice records (1990). We discard the first 10 years of the integrations are maintained by the Icelandic Meteorological Office to avoid the initial shock in changing the wind forcing (Wallevik and Sigurjónsson 1998). During the from the climatological values, and we analyze the pe- “GSA’70” event, anomalous sea ice exported from the riod of 1956–90. To compare with observations, we de- Arctic entered into the North Atlantic through the fine the “climatology” as the average for the 35-yr pe- Denmark Strait off the Icelandic coast (Dickson et al. riod of 1956–90. All anomalous responses shown below 1988). Walsh and Chapman (1990) used the Koch index are relative to the defined climatology. as a proxy for the GSA events. Similarly, the GSA b. Model results index we used in the simulations correlates with the Iceland sea ice extent anomaly. We do note that the We found that, typically, anomalous radiative forcing GSAs are not necessarily caused solely by the anoma- due to increased greenhouse gases has little impact on lous freshwater and sea ice from the Arctic via the the North Atlantic decadal variations as compared with . Local forcing over the Labrador Sea re- that of other forcings. For the experiment forced only gion is important in causing the “GSA’80” and with the COADS anomalous wind forcing and the “GSA’90” events as shown in Belkin et al. (1998) and anomalous radiative forcing due to increased green- Belkin (2004). Hence, the Iceland sea ice extent house gases (GW ϩ WIND), that is, the anomalous anomaly underestimates the strength of GSA’80 and freshwater flux is not included, the modeled anomalous GSA’90. freshwater export from the Arctic through the Fram Walsh and Chapman (1990) also suggested that GSA Strait is very small and is negligible when compared events are correlated with the difference of SLP with that found in Häkkinen (1993). This is mainly be- anomaly between the Greenland and the Arctic. The cause the COADS anomalous wind forcing does not GSA index used in our simulations represents both the contain adequate data over the Arctic region, and thus remote forcing from the Arctic north of the Kara Sea the experiment cannot simulate well the wind-driven and the local forcing from the Labrador Sea off the freshwater transport through the Fram Strait. Hence, Greenland west coast for GSA events. Its amplitude is the above experiment (GW ϩ WIND) cannot obtain based on the estimated anomalous freshwater flux en- the GSA forcing, and we included the anomalous fresh- tering the North Atlantic during the GSA’70 (Dickson water flux (Fig. 2a) in experiments (GSA and “GW ϩ et al. 1988). We focused on the GSA’70 and GSA’80 WIND ϩ GSA”) to represent the total anomalous events (positive phases around late 1960s and late 1970s freshwater flux entering into the North Atlantic as a in the GSA index and observed Iceland sea ice extent result of GSA events. Note that we do not try to model anomaly; Fig. 2a); no anomalous freshwater flux is ap- the root cause of GSA events; rather, we impose the plied after 1984. anomalous freshwater flux entering into the North At- We perform several perturbed experiments (listed in lantic as a result of GSA events. Table 1) with various combinations of the anomalous With all anomalous forcings (anomalous radiative external forcing, including those due to the increase of forcing due to increased greenhouse gases, COADS

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC 474 JOURNAL OF CLIMATE VOLUME 19

TABLE 2. Modeled decadal anomalies.

Modeled decadal anomalies GW ϩ WIND GSA GW ϩ WIND ϩ GSA Labrador Sea SST anomaly in 1970 (K) 0.04 Ϫ0.45 Ϫ0.62 Labrador Sea SST anomaly in 1983 (K) Ϫ0.17 Ϫ1.23 Ϫ1.57 Max DWBC weakening in the 1970s (Sv) 1.3 1.0 2.0 Max DWBC weakening in the 1980s (Sv) 1.0 1.6 2.6 Max NRG weakening in the 1970s (Sv) 4.0 2.7 6.3 Max NRG weakening in the 1980s (Sv) 2.5 4.7 9.6 Max northward Gulf Stream shift in the 1970s (°) 0.37 0.22 0.58 Max northward Gulf Stream shift in the 1980s (°) 0.29 0.24 0.76

anomalous wind forcing, and the anomalous freshwater mum of 4.0 Sv in 1974 and 2.5 Sv in 1986, and the flux forcing due to GSA events; GW ϩ WIND ϩ GSA) modeled mean Gulf Stream path shifts northward by a together, the modeled SST anomaly in the Labrador maximum of 0.37° in 1976 and 0.29° in 1990 (Table 2). Sea shows significant cooling in the early 1970s and These responses are mainly due to the reduced wind early 1980s; for example, the anomalous cooling speed over the Labrador Sea during negative NAO reaches 0.62 K in 1970 and 1.57 K in 1983, respectively, phases at 1970 and 1980, respectively, which induces a similar to the observed cooling [Fig. 2b; Hadley Centre reduction of outgoing surface heat flux, a weak surface Sea Ice and Sea Surface Temperature (HadISST) warming, and weakened deep-water formation, result- dataset; Rayner et al. 2003]. For the experiment (GSA) ing in the weakening of THC, DWBC, and NRG and forced only with the anomalous freshwater flux due to northward shifts of Gulf Stream path several years GSA events (Fig. 2a), the modeled SST anomaly in the later. The increased greenhouse gases also contribute Labrador Sea shows cooling of 0.45 K in 1970 and 1.23 to the long-term weakening trend of the THC/DWBC K in 1983, respectively (Table 2). For the experiment in the above experiment (GW ϩ WIND). For example, (GW ϩ WIND) forced only with the COADS anoma- for the experiment (“GW”) forced with only the lous wind forcing and the anomalous radiative forcing anomalous radiative forcing due to increased green- due to increased greenhouse gases, the modeled SST house gases, the DWBC is linearly reduced by 0.86 Sv anomaly in the Labrador Sea shows 0.04 K warming in from 1956 to 1990. For the experiment (GSA) forced 1970 and 0.17 K cooling in 1983, respectively (Table 2). only with the anomalous freshwater flux due to GSA Hence, the modeled Labrador Sea cooling in the early events (Fig. 2a), the modeled DWBC is weakened by a 1970s and early 1980s is mainly due to GSA’70 and maximum of 1.0 Sv in 1973 and 1.6 Sv in 1985, the GSA’80 events. modeled NRG is weakened by a maximum of 2.7 Sv in With all anomalous forcings (GW ϩ WIND ϩ GSA), 1974 and 4.7 Sv in 1987, and the modeled mean Gulf the modeled DWBC near the Grand Banks, cyclonic Stream path shifts northward by a maximum of 0.22° in NRG (Figs. 2c,d), and maximum THC are weakened in 1976 and 0.24° in 1989 (Table 2). the mid-1970s and mid- to late 1980s. The modeled The modeled northward shifts of the mean Gulf mean Gulf Stream path shifts to the north in the mid- Stream path (Fig. 2e) lag the weakening of NRG (Fig. 1970s and around 1990, which is also shown in diag- 2d) by 1 yr at the maximum correlation (r ϭ 0.87) be- noses from the World Ocean Database 1998 (WOD98; cause of the westward propagation of Rossby waves Levitus et al. 1998; Fig. 2e). The modeled total weak- excited by DWBC variations south of Grand Banks and ening of NRG is consistent with interpentadal varia- thus may lag the warming off the U.S. east coast in- tions of NRG diagnosed from observations (Great- duced by the weakening of the cyclonic NRG. Indeed, batch et al. 1991; Ezer et al. 1995). The contributions both observation (WOD98) and modeling results (Fig. from GSA events and the anomalous wind forcing to 3) show that a large-scale dipole pattern (warming off the total weakening of DWBC, THC, and NRG and the U.S. east coast and cooling south of Greenland) northward shifts of mean Gulf Stream path are on the leads the northward shift of the mean Gulf Stream path same order. For the experiment (GW ϩ WIND) forced by about 1 yr. This dipole pattern with the warming only with the COADS anomalous wind forcing and the confined to NRG region is different from typical NAO- anomalous radiative forcing due to increased green- forced tripolar pattern (Kushnir et al. 2002) induced house gases, the modeled DWBC is weakened by a directly by the wind and the low-frequency NAO- maximum of 1.3 Sv (1 Sv ϵ 106 m3 sϪ1) in 1972 and 1.0 forced one-sign monopole pattern of SST anomaly Sv in 1984, the modeled NRG is weakened by a maxi- (Visbeck et al. 1998), indicating that it is not only forced

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 475

FIG. 3. Normalized covariance between SST anomaly and 1-yr lagged shifts of Gulf Stream path (1956–90). (a) Diagnosed from observations (WOD98). (b) Results from the numerical experiment with all anomalous forcings (GW ϩ WIND ϩ GSA). The difference between the integration of covariance over the domain off the U.S. east coast (38°–43°N, 70°–52°W) and over the domain south of Greenland (48°–60°N, 48°–38°W) reaches the maximum when shifts of Gulf Stream path lagged the SST anomaly by about 1 yr in both observation and modeling results. by the wind. The cooling south of Greenland is mainly NAO at low frequency, and in particular can efficiently due to the propagation of low-salinity anomaly in the force the low-frequency positive NAO phases. This is subpolar gyre. The warming off the U.S. east coast is by no means the universal response of AGCMs, and not forced by air–sea heat flux, but is due to variations other modeling studies with AGCMs show a weak re- of oceanic advection associated with the weakening of sponse of NAO to extratropical SST forcing (Kushnir NRG. The atmosphere provides negative feedback to et al. 2002). However, if the dipolar SST anomaly pat- this dipole SST anomaly; that is, more surface heat flux tern seen in both observations and our numerical re- enters into the subpolar North Atlantic, and less sur- sults (warming off the U.S. east coast and cooling south face heat flux is released into the atmosphere off the of Greenland; Fig. 3) can force positive NAO phases at U.S. east coast, indicating a reduced poleward oceanic low frequency, then GSA events may lead to positive heat transport, consistent with the weakening of THC. phases of NAO at low frequency by triggering just such The differences between the modeled and observed di- SST anomaly patterns. As noted, this effect is very dif- pole pattern (Fig. 3) are mainly due to the model biases ficult to reliably model numerically, so we test it with in the mean state ocean circulation. For example, in the analyses of three independent observed time series of modeled mean state, the part of NRG south of New- the twentieth century: (i) Iceland sea ice extent foundland is too weak compared to observations; thus anomaly, derived from one of the long-term Icelandic the warming there associated with the weakening of the sea ice records, the number of areas off the coasts of NRG is very small (Fig. 3b), while the observed warm- Iceland on which sea ice is observed in a sea ice year ing off the U.S. east coast extends farther eastward to (Wallevik and Sigurjónsson 1998), by removing the south of Newfoundland (Fig. 3a); the modeled cooling trend for the whole period; (ii) Labrador Sea SST south of Greenland does not extend as wide as that anomaly (HadISST; Rayner et al. 2003); (iii) the NAO observed (Fig. 3), because the modeled mean state sub- index (Hurrell 1995). At decadal or longer time scales, polar gyre circulation is weaker than observed. the positive GSA events (more Iceland sea ice extent) occurred in the early twentieth century as well as in the late 1960s and late 1970s, and the negative GSA phase 3. Observed relationship between GSA and NAO (less Iceland sea ice extent) occurred in the middle Some simulations with atmospheric GCMs (Rodwell twentieth century (Fig. 4a). Both the Labrador Sea sur- et al. 1999; Mehta et al. 2000) suggest that a dipolar SST face cooling (warming) and positive (negative) NAO anomaly pattern (warming off the U.S. east coast and phase (Figs. 4b,c) lagged the positive (negative) GSA cooling south of Greenland) can modulate the phase of events by a few years, respectively. The negative NAO

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 3 live 4/C 476 JOURNAL OF CLIMATE VOLUME 19

FIG. 4. Observed time series during the twentieth century. (a) Annual Iceland sea ice extent anomaly; the original dataset is maintained by the Icelandic Meteorological Office (Wallevik and Sigurjónsson 1998). (b) Central Labrador Sea (60°N, 55°W) annual mean (inverted) SST anomaly from HadISST dataset (K). (c) Winter (December–March) principal-component- based NAO index (Hurrell 1995, provided by the Climate Analysis Section, NCAR, Boulder, CO). The green dashed line is the unfiltered data, and the color shading is the 8-yr low-pass- filtered data. The linear trends for the whole period (1901–90) of the time series are removed. trend in the middle twentieth century (before the 1970s; that is, the positive GSA index leads modeled Labrador Fig. 4c) cannot be explained by global warming (Hoer- Sea surface cooling by 3 yr (r ϭϪ0.66), the approxi- ling et al. 2001) or stratospheric ozone depletion (Hart- mate propagation time scale of the salinity anomaly mann et al. 2000), and may have been triggered by the from east Greenland coast to central Labrador Sea. negative GSA phase several years prior. The positive GSA index also leads the modeled weak- The lag relationship between NAO and GSA events ening of NRG (Fig. 5b) by 6 yr (r ϭ 0.78). The modeled at low frequency is robust throughout the whole period Labrador Sea surface cooling leads the weakening of and suggests that there may be some long-term predict- NRGby4yr(r ϭϪ0.7), that is, the advection time ability. Indeed at low frequency (Fig. 5a) the positive scale of DWBC from the Labrador Sea to south of Iceland sea ice extent anomaly leads the positive NAO the Grand Banks. There is almost no lag between phase by 7 yr at the maximum cross correlation (r ϭ modeled NRG anomaly and dipole SST anomaly, and if 0.66) and leads the Labrador Sea surface cooling by 3 yr the dipole SST anomaly can modulate the low- (r ϭϪ0.85); the Labrador Sea surface cooling leads the frequency NAO phases (Rodwell et al. 1999), this positive NAO phase by 4 yr (r ϭϪ0.64). Simulated might provide a mechanism for the observed phase re- variables show similar phase relationships (Fig. 5b); lationships (Fig. 5a).

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 4 live 4/C 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 477

FIG. 5. Cross correlation and cross-spectral analysis of observed variables and comparisons with some modeling results. (a) Cross correlations between low-pass-filtered observed variables shown in Fig. 4. Iceland SI: Iceland sea ice extent anomaly; LS SST: Labrador Sea SST anomaly. (b) Cross correlations between unfiltered modeled variables shown in Fig. 2. NRG (inv): inverted NRG anomaly. (c) Squared coherence between unfiltered observed variables, i.e., the green dashed lines in Fig. 4. The horizontal dashed line is the 95% confidence level. (d) Time lag in years between unfiltered observed variables, i.e., the green dashed lines in Fig. 4. Negative value means that variable 1 leads variable 2. LS SST (inv): inverted Labrador Sea SST anomaly.

Strong coherence (significant at the 95% level) be- anomaly leads the Labrador Sea surface cooling (warm- tween any two of the above unfiltered observed vari- ing) by an average of about 3 yr; the Labrador Sea ables appears at decadal and multidecadal time scales surface cooling (warming) leads the positive (negative) (Fig. 5c). For example, the squared coherence between NAO phase by an average of about 4 yr; the positive the unfiltered Iceland sea ice extent anomaly and the (negative) Iceland sea ice extent anomaly leads the NAO index is about 0.59 at decadal time scales (0.1 positive (negative) NAO phase by an average of about cycles per year; Fig. 5c) and even higher at longer time 7 yr (Fig. 5d), consistent with that shown in the Fig. 5a. scales (the 95% significance level is 0.527). The squared At interannual time scales, the positive (negative) coherence between the unfiltered Labrador Sea SST NAO phase is nearly in phase with the Labrador Sea anomaly and NAO index is slightly stronger, and the surface cooling (warming) and the Iceland sea ice ex- squared coherence between the unfiltered Iceland sea tent decrease (increase) (Fig. 5d), that is, the ocean is ice extent anomaly and the Labrador Sea SST anomaly passively forced by NAO generated by internal atmo- is even much stronger at the low frequency (Fig. 5c). spheric variabilities (Vallis et al. 2004). At low frequen- The cross spectral analysis of unfiltered observed vari- cies the NAO phases are more likely to be influenced ables also shows that at decadal and multidecadal time by oceanic processes, consistent with a previous study scales, the positive (negative) Iceland sea ice extent that the correlation between the simulated NAO index

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 5 live 4/C 478 JOURNAL OF CLIMATE VOLUME 19

FIG. 6. Comparisons with observed Barents/ and Labrador Sea winter sea ice anoma- lies. (a) Annual Iceland sea ice extent anomaly; the original dataset is maintained by the Icelandic Meteorological Office (Wallevik and Sigurjónsson 1998). (b) Observed Barents/Greenland Sea (70°– 77°N, 10°W–50°E) winter (DJF) sea ice concentration anomaly, plotted in the year in which January occurs. (c) Observed central Labrador Sea (60°N, 55°W) annual mean (inverted) SST anomaly from HadISST dataset (K). (d) Observed Labrador Sea (56°–65°N, 65°–50°W) winter (DJF) sea ice concen- tration anomaly, plotted in the year in which January occurs. Green dashed line: (a), (c) unfiltered; (b), (d) normalized anomaly smoothed with a three-point binomial filter [for (b) adapted from Fig. 2 and for (d) adapted from Fig. 3 in Deser and Timlin 1996]. Color shading: (a)–(d) 8-yr low-pass filtered.

(forced by observed SST anomaly) and observed NAO ter sea ice concentration anomaly (Fig. 6b), with a lead index is much higher for low-pass-filtered data than of 0–1 yr. The slight time lead is probably because the unfiltered data (Mehta et al. 2000). Iceland sea ice extent anomaly is for annual mean, and Similar to the Koch index, the Iceland sea ice extent the Barents/Greenland Sea winter sea ice anomaly is anomaly we derived from one of the long-term Icelan- only for winter [December–February (DJF)]. At low dic sea ice records is a rough estimation of the sea ice frequency, the increase of Barents/Greenland Sea win- severity off the Icelandic coast. To further verify the ter sea ice cover leads the Labrador Sea surface cooling observed datasets discussed above, we compared them (Fig. 6c) by about 2–3yr(r ϭ 0.81) and leads the posi- with observations from another independent sea ice tive NAO phase by about 5–6yr(r ϭ 0.49), consistent data source: the observed Barents/Greenland Sea win- with the relationship shown in Fig. 5. In the Labrador ter sea ice concentration anomaly (taken from Fig. 2 in Sea, surface cooling conditions are associated with high Deser and Timlin 1996), and the observed Labrador sea ice concentrations (Figs. 6c,d). At low frequency, Sea winter sea ice concentration anomaly (taken from the observed Labrador Sea SST anomaly we discussed Fig. 3 in Deser and Timlin 1996; also shown in Belkin et before (Fig. 6c) has very high negative correlation (r ϭ al. 1998). We found that at low frequency, the Iceland Ϫ0.96) with the observed Labrador Sea winter sea ice sea ice extent anomaly (Fig. 6a) is highly correlated (r anomaly (Fig. 6d), with a lead of 1 yr. The slight time ϭ 0.86) with the observed Barents/Greenland Sea win- lead is probably because the SST anomaly is for annual

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC

Fig 6 live 4/C 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 479 mean, and the sea ice anomaly is only for winter (DJF). dictable atmospheric noise, then the system would have At low frequency, the increase of Iceland sea ice extent no long-term predictability (e.g., Bretherton and Bat- leads the increase of the Labrador Sea winter sea ice tisti 2000). However, if the SST anomaly were to be cover by about 4.5 yr (r ϭ 0.81), and the increase of the triggered by GSA events several years in advance, Labrador Sea winter sea ice cover leads the positive there would be a degree of long-term predictability to NAO phase by about 2.5 yr (r ϭ 0.67). This is also the system. An analysis of long-term observations sug- consistent with the relationship shown in Fig. 5. gests that, in fact, the latter may be the case: on such Deser and Timlin (1996) mentioned the observa- long time scales, the positive (negative) Iceland sea ice tional evidence that for both of the decadal events in 1) extent anomaly leads Labrador Sea surface cooling the late 1960s–early 1970s and 2) late 1970s–early (warming) and the positive (negative) NAO phase by 1980s, there was systematic movement of the region of about 3 and 7 yr, respectively, suggesting that the high sea ice concentrations through Denmark Strait phases of NAO at decadal and multidecadal time scales around the southern tip of Greenland and into the may be influenced by oceanic processes. Labrador Sea, and high sea ice concentrations in Bar- To investigate the oceanic mechanisms involved, we ents/Greenland Sea were followed 3–4 yr later by high performed various modeling studies, and these show sea ice concentrations in the Labrador Sea. These are that positive GSA phase typically leads the Labrador consistent with the above analysis. A recent observa- Sea surface cooling by about 3 yr and leads the weak- tional analysis (Deser et al. 2002) also showed that the ening of NRG, which is associated with the dipolar SST freshwater anomalies at 100-m depth in the West anomaly, by about 6 yr. The dipole SST anomaly Greenland Current preceded the decadal sea ice cover (warming off the U.S. east coast and cooling south of anomaly in the northern Labrador Sea by about 8 Greenland) associated with the positive GSA phase, is months. Note that the positive peak (Figs. 6a,b) in the consistent with a reduction of THC and poleward oce- late 1970s to early 1980s was weaker than that in the anic heat transport. The atmosphere may have to trans- late 1960s to early 1970s in both the Iceland sea ice port more heat poleward to compensate for the re- extent anomaly and the Barents/Greenland Sea winter duced oceanic poleward heat transport, and it is pos- sea ice concentration anomaly; whereas the positive sible that this may lead to more storms and a positive peak (Figs. 6c,d) in the early 1980s is as large as or even NAO phase. It is not possible to simulate the impact of stronger than that in the early 1970s in both the Labra- the dipolar SST anomalies on NAO with the EBM that dor Sea SST anomaly and the Labrador Sea winter sea we used to model the atmosphere, and the precise ice concentration anomaly. This is probably because of mechanism whereby oceanic variability affects the at- the contribution of local forcing to the GSA’80 event mospheric variability on long time scales is beyond our (Belkin et al. 1998), and thus the Iceland sea ice extent current scope. The NAO may provide negative feed- anomaly and Barents/Greenland Sea winter sea ice backs to GSA events, for example, the decadal oscilla- concentration anomaly underestimate the strength of tions after 1970 may be caused by coupled interactions the GSA’80 event. between NAO and GSAs. Let us now briefly discuss some of the data used and 4. Conclusions and discussion limitations of our study. The analyzed low-frequency relationship between Barents/Greenland Sea winter sea GSA events can evidently play a very important role ice (Labrador Sea winter sea ice anomaly) and the in causing large-scale coherent low-frequency variabili- NAO index is based on time series of a relatively short ties in the North Atlantic, such as the weakening of period (1955–90). Observed sea ice data are sparse be- THC, DWBC, and NRG and the northward shifts of fore the 1950s, and satellite-based sea ice observations mean Gulf Stream path. A large-scale North Atlantic are only available after the 1970s. To study the low- dipolar SST anomaly (warming off the U.S. east coast frequency North Atlantic variability, especially at the and cooling south of Greenland), which leads north- multidecadal time scales, we ideally need very long ward shifts of mean Gulf Stream path by about 1 yr, term observed time series. The records of both the ob- appears in both observations and our modeling results. servational-based long-term Iceland sea ice extent Such an anomaly is important because some modeling anomaly (which is highly correlated with Barents/ studies (e.g., Rodwell et al. 1999) suggest that it may Greenland Sea winter sea ice anomaly during 1955–90) modulate the phases of NAO at low frequency, al- and the Labrador Sea SST anomaly (which has very though it should also be said that the physical processes strong negative correlation with the Labrador Sea win- that control the amplitudes of NAO are not well un- ter sea ice anomaly during 1955–90) go well back before derstood. If this SST anomaly were driven by unpre- the 1950s; they both indicate that GSA events might

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC 480 JOURNAL OF CLIMATE VOLUME 19 have occurred during the early twentieth century and tive NAO trend before the late 1960s with a lead of that they both lead the NAO phases by several years on several years, respectively. The negative GSA phase decadal and multidecadal time scales. Dickson et al. during the middle twentieth century might have con- (1988) suggested that the fragmentary ocean data indi- tributed to the strengthening of THC and the warm cate low-salinity events in the North Atlantic during the AMO phase. early twentieth century. The low-frequency variations In summary, our modeling results suggest that the of the Iceland sea ice extent anomaly (Fig. 4a) are also impact of GSAs (both positive and negative phases), consistent with a recent reconstruction of the Fram which are often missing in model simulations of North Strait sea ice export (Schmith and Hansen 2003). They Atlantic low-frequency variability, are at least as im- found that GSAs observed around 1968–70 and 1980– portant as the impact of the observed anomalous wind 82 both occurred when the Fram Strait sea ice export forcing for the production of large-scale SST anomalies. was high; prior to these there was a long period with The results also suggest that, to simulate the low- low Fram Strait sea ice export and no GSAs, but early frequency temperature variations off the U.S. east in the nineteenth century several GSAs occurred coast, it is very important for climate models to prop- (Schmith and Hansen 2003). Nevertheless, sea ice ex- erly simulate the mean state NRG and the variations of tent anomalies along the Icelandic coast may not be NRG in response to the variations of DWBC. Of always associated with the Arctic sea ice export anoma- course, the ultimate origin of GSA events is still not lies. They can be driven by local processes not related very clear. They may originate from coupled interac- to sea ice export anomalies from the Arctic, and they tions of Arctic sea ice, river runoff, and atmospheric more likely reflect the anomalous sea ice/freshwater and oceanic processes (Mysak et al. 1990), and local entering into the North Atlantic through the Denmark processes over the Labrador Sea are also very impor- Strait. tant (Belkin et al. 1998; Belkin 2004). Evidently, under- Deser and Timlin (1996) showed that the observed standing and simulating proper GSA events, and their high winter sea ice conditions in the Labrador Sea pre- potential impact on the climate system at large, remains ceded the surface cooling in the North Atlantic subpo- a challenge for climate modeling efforts. lar region, consistent with our modeling results that GSA events induce widespread cooling south of Green- Acknowledgments. We thank Thor Jakobsson at the land. The multidecadal variations of the Iceland sea ice Icelandic Meteorological Office for providing the ob- extent and Labrador Sea SST (Fig. 4) might have con- servational dataset of the Icelandic Sea Ice records. We tributed to the Atlantic multidecadal oscillation also thank Isaac Held for initially providing us with the (AMO). The AMO has cool phases in the North At- EBM, Mike Winton for providing us with the sea ice lantic during 1905–25 and 1970–90, a warm phase in the model, and many GFDL colleagues for the develop- North Atlantic in the middle twentieth century, and it is ment of MOM4 and the Flexible Modeling System thought to be linked to variations in the THC (Del- (FMS) used in this study. We thank Igor Belkin and an worth and Mann 2000; Enfield et al. 2001). It has been anonymous reviewer for very helpful comments on this shown that the AMO has significant impacts on the paper. The numerical experiments were carried with multidecadal climate variability over North America the supercomputer facilities at NOAA/GFDL, and the and Europe (Sutton and Hodson 2005), and the mul- work was partially funded by the NSF Ocean Science tidecadal variability of the Atlantic hurricane activity Division. (Goldenberg et al. 2001). The observed GSA events often refer to positive GSA phases, that is, more sea REFERENCES ice/freshwater and low salinity conditions in the North Atlantic subpolar region. The positive GSA phases that Belkin, I. M., 2004: Propagation of the “Great Salinity Anomaly” occurred in the early twentieth century as well as those of the 1990s around the northern North Atlantic. Geophys. Res. Lett., 31, L08306, doi:10.1029/2003GL019334. after the 1960s might have contributed to the weaken- ——, S. Levitus, J. Antonov, and S. Malmberg, 1998: ‘Great sa- ing of THC and the cool AMO phases. The negative linity anomalies’ in the North Atlantic. Progress in Oceanog- GSA phases—that is, periods with less sea ice/ raphy, Vol. 41, Pergamon, 1–68. freshwater and high-salinity conditions in the North At- Bretherton, C. S., and D. S. Battisti, 2000: An interpretation of the lantic subpolar region—are also very important for the results from atmospheric general circulation models forced North Atlantic low-frequency variability. For example, by the time history of the observed sea surface temperature distribution. Geophys. Res. Lett., 27, 767–770. during the middle twentieth century, the long-term Curry, R. G., M. S. McCartney, and T. M. Joyce, 1998: Oceanic negative GSA phase might have contributed to the transport of subpolar climate signals to mid-depth subtropical long-term Labrador Sea surface warming and the nega- waters. Nature, 391, 575–577.

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC 1FEBRUARY 2006 NOTES AND CORRESPONDENCE 481 da Silva, A., A. C. Young, and S. Levitus, 1994: Algorithms and Joyce, T. M., C. Deser, and M. A. Spall, 2000: The relation be- Procedures. Vol. 1, Atlas of Surface Marine Data 1994, tween decadal variability of subtropical mode water and the NOAA Atlas NESDIS 6, 83 pp. North Atlantic Oscillation. J. Climate, 13, 2550–2569. Delworth, T. L., and M. E. Mann, 2000: Observed and simulated Kelly, P. M., C. M. Goodess, and B. S. G. Cherry, 1987: The in- multidecadal variability in the Northern Hemisphere. Cli- terpretation of the Icelandic sea ice record. J. Geophys. Res., mate Dyn., 16, 661–676. 92, 10 835–10 843. ——, S. Manabe, and R. J. Stouffer, 1993: Interdecadal variations of the thermohaline circulation in a coupled ocean– Kushnir, Y., 1994: Interdecadal variations in North Atlantic sea atmosphere model. J. Climate, 6, 1993–2011. surface temperature and associated atmospheric conditions. Deser, C., and M. L. Blackmon, 1993: Surface climate variations J. Climate, 7, 141–157. over the North Atlantic ocean during winter: 1900–1989. J. ——, W. A. Robinson, I. Bladé, N. M. J. Hall, S. Peng, and R. T. Climate, 6, 1743–1753. Sutton, 2002: Atmospheric GCM response to extratropical ——, and M. S. Timlin, 1996: Decadal variations in sea ice and sea SST anomalies: Synthesis and evaluation. J. Climate, 15, surface temperature in the subpolar North Atlantic. The At- 2233–2256. lantic Climate Change Program, Proc. from the Principal In- Latif, M., 2001: Tropical Pacific/Atlantic ocean interactions at vestigators Meeting, Woods Hole, MA, WHOI, 36–40. multi-decadal time scales. Geophys. Res. Lett., 28, 539–542. ——, M. Holland, G. Reverdin, and M. Timlin, 2002: Decadal Lazier, J. R. N., 1980: Oceanographic conditions at Ocean variations in Labrador Sea ice cover and North Atlantic sea Weather Ship Bravo, 1964–74. Atmos.–Ocean, 18, 227–238. surface temperatures. J. Geophys. Res., 107, 3035, doi:10.1029/2000JC000683. Levitus, S., and Coauthors, 1998. Introduction. Vol. 1, World Dickson, R. R., J. Meincke, S.-A. Malmberg, and A. J. Lee, 1988: Ocean Database 1998 (WOD98), 346 pp. The ‘Great Salinity Anomaly’ in the northern North Atlantic Manabe, S., K. Bryan, and M. J. Spelman, 1990: Transient re- 1968–1982. Progress in Oceanography, Vol. 20, Pergamon, sponse of a global ocean–atmosphere model to a doubling of 103–151. atmospheric carbon dioxide. J. Phys. Oceanogr., 20, 722–749. Eden, C., and T. Jung, 2001: North Atlantic interdecadal variabil- Mehta, V. M., M. J. Suarez, J. V. Manganello, and T. L. Delworth, ity: Oceanic response to the North Atlantic Oscillation 2000: Oceanic influence on the North Atlantic Oscillation (1865–1997). J. Climate, 14, 676–691. and associated Northern Hemisphere climate variations: Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The 1959–1993. Geophys. Res. Lett., 27, 121–124. Atlantic multidecadal oscillation and its relation to rainfall Mysak, L. A., D. K. Manak, and D. F. Marsden, 1990: Sea-ice and river flows in the continental U.S. Geophys. Res. Lett., 28, 2077–2080. anomalies observed in the Greenland and Labrador Seas dur- ing 1901–1984 and their relations to an interdecadal Arctic Ezer, T., G. L. Mellor, and R. J. Greatbatch, 1995: On the inter- climate cycle. Climate Dyn., 5, 111–133. pentadal variability of the North Atlantic Ocean: Model simulated changes in transport, meridional heat flux and Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. coastal sea level between 1955–1959 and 1970–1974. J. Geo- Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: phys. Res., 100, 10 559–10 566. Global analyses of sea surface temperature, sea ice, and night Gerdes, R., and C. Köberle, 1995: On the influence of DSOW in marine air temperature since the late nineteenth century. J. a numerical model of the North Atlantic general circulation. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670. J. Phys. Oceanogr., 25, 2624–2641. Robertson, A. W., C. R. Mechoso, and Y.-J. Kim, 2000: The in- Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nuñez, and fluence of sea surface temperature anomalies on the North W. M. Gray, 2001: The recent increase in Atlantic hurricane Atlantic Oscillation. J. Climate, 13, 122–138. activity: Causes and implications. Science, 293, 474–479. Rodwell, M. J., D. P. Rowell, and C. K. Folland, 1999: Oceanic Greatbatch, R. J., A. F. Fanning, A. D. Goulding, and S. Levitus, forcing of the wintertime North Atlantic Oscillation and Eu- 1991: A diagnosis of interpendatal circulation changes in the ropean climate. Nature, 398, 320–323. North Atlantic. J. Geophys. Res., 96, 22 009–22 023. Griffies, S. M., M. J. Harrison, R. C. Pacanowski, and A. Rosati, Schmith, T., and C. Hansen, 2003: Fram Strait ice export during 2004: A technical guide to MOM4. GFDL Ocean Group the nineteenth and twentieth centuries reconstructed from a Tech. Rep. 5, NOAA/GFDL, 339 pp. multiyear sea ice index from southwestern Greenland. J. Cli- Häkkinen, S., 1993: An Arctic source for the Great Salinity mate, 16, 2782–2791. Anomaly: A simulation of the Arctic ice-ocean system for Sutton, R. T., and D. L. R. Hodson, 2005: Atlantic ocean forcing 1955–1975. J. Geophys. Res., 98, 16 397–16 410. of North American and European summer climate. Science, Halliwell, G. R., Jr., 1997: Decadal and multidecadal North At- 309, 115–118. lantic SST anomalies driven by standing and propagating ba- Taylor, A. R., and J. A. Stephens, 1998: The North Atlantic os- sin-scale atmospheric anomalies. J. Climate, 10, 2405–2411. cillation and the latitude of the Gulf Stream. Tellus, 50, 134– Hartmann, D. L., J. M. Wallace, V. Limpasuvan, D. W. J. Thomp- 142. son, and J. R. Holton, 2000: Can ozone depletion and green- Thompson, J. D., and W. J. Schmitz, 1989: A limited-area model house warming interact to produce rapid climate change? of the Gulf Stream: Design, initial experiments, and model– Proc. Natl. Acad. Sci. USA, 97, 1412–1417. data intercomparison. J. Phys. Oceanogr., 19, 791–814. Hoerling, M. P., J. W. Hurrell, and T. Xu, 2001: Tropical origins for recent North Atlantic climate change. Science, 292, 90–92. Vallis, G. K., E. Gerber, P. Kushner, and B. Cash, 2004: A mecha- Hurrell, J. W., 1995: Decadal trends in the North Atlantic Oscil- nism and simple model of the North Atlantic Oscillation and lation: Regional temperatures and precipitation. Science, 269, annular modes. J. Atmos. Sci., 61, 264–280. 676–679. Visbeck, M., H. Cullen, G. Krahmann, and N. Naik, 1998: An

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC 482 JOURNAL OF CLIMATE VOLUME 19

ocean model’s response to North Atlantic Oscillation-like Weaver, A. J., J. Marotzke, P. F. Cummins, and E. S. Sarachik, wind forcing. Geophys. Res. Lett., 25, 4521–4524. 1993: Stability and variability of the thermohaline circulation. Wallevik, J. E., and H. Sigurjónsson, 1998. The Koch Index: For- J. Phys. Oceanogr., 23, 39–60. mulation, corrections and extension. Icelandic Meteorologi- ——, and Coauthors, 2001: The UVic Earth System Climate cal Office Rep. VÍ-G98035-ÚR28, 14 pp. Model: Model description, climatology and application to Walsh, J. E., and W. L. Chapman, 1990: Arctic contribution to past, present and future climates. Atmos.–Ocean, 39, 361–428. upper-ocean variability in the North Atlantic. J. Climate, 3, Winton, M., 2000: A reformulated three-layer sea ice model. J. 1462–1473. Atmos. Oceanic Technol., 17, 525–531.

Unauthenticated | Downloaded 10/03/21 08:37 PM UTC